Merge pull request #3145 from harlouci/props_tensor_nnet

theano/tensor/nnet/Conv3D.py

@@ -45,14 +45,7 @@ class Conv3D(theano.Op):
     """ 3D `convolution` of multiple filters on a minibatch
         :note: does not flip the kernel, moves kernel with a user specified stride
     """
-    def __eq__(self, other):
-        return type(self) == type(other)
-
-    def __hash__(self):
-        return hash(type(self))
-
-    def __str__(self):
-        return "Conv3D"
+    __props__ = ()
 
     def c_code_cache_version(self):
         return (3, blas_header_version())

theano/tensor/nnet/ConvGrad3D.py

@@ -12,11 +12,7 @@ from theano.gradient import DisconnectedType
 
 class ConvGrad3D(theano.Op):
     """ Gradient of Conv3D with respect to W """
-    def __eq__(self, other):
-        return type(self) == type(other)
-
-    def __hash__(self):
-        return hash(type(self))
+    __props__ = ()
 
     def c_code_cache_version(self):
         return (1,)

theano/tensor/nnet/ConvTransp3D.py

@@ -10,11 +10,7 @@ from theano.gradient import DisconnectedType
 
 class ConvTransp3D(theano.Op):
     """ "Transpose" of Conv3D (Conv3D implements multiplication by an implicitly defined matrix W. This implements multiplication by its transpose) """
-    def __eq__(self, other):
-        return type(self) == type(other)
-
-    def __hash__(self):
-        return hash(type(self))
+    __props__ = ()
 
     def c_code_cache_version(self):
         return (3,)

theano/tensor/nnet/conv3d2d.py

@@ -73,6 +73,8 @@ class DiagonalSubtensor(Op):
     see what's necessary at that point.
 
     """
+    __props__ = ("inplace",)
+
     def __str__(self):
         if self.inplace:
             return "%s{inplace}" % self.__class__.__name__
@@ -83,12 +85,6 @@ class DiagonalSubtensor(Op):
         if inplace:
             self.view_map = {0: [0]}
 
-    def __eq__(self, other):
-        return type(self) == type(other) and self.inplace == other.inplace
-
-    def __hash__(self):
-        return hash((type(self), self.inplace))
-
     def make_node(self, x, i0, i1):
         _i0 = tensor.as_tensor_variable(i0)
         _i1 = tensor.as_tensor_variable(i1)
@@ -117,6 +113,8 @@ class IncDiagonalSubtensor(Op):
     """
     The gradient of DiagonalSubtensor
     """
+    __props__ = ("inplace",)
+
     def __str__(self):
         if self.inplace:
             return "%s{inplace}" % self.__class__.__name__
@@ -127,12 +125,6 @@ class IncDiagonalSubtensor(Op):
         if inplace:
             self.destroy_map = {0: [0]}
 
-    def __eq__(self, other):
-        return type(self) == type(other) and self.inplace == other.inplace
-
-    def __hash__(self):
-        return hash((type(self), self.inplace))
-
     def make_node(self, x, i0, i1, amt):
         _i0 = tensor.as_tensor_variable(i0)
         _i1 = tensor.as_tensor_variable(i1)

theano/tensor/nnet/neighbours.py

@@ -13,6 +13,9 @@ import numpy
 
 
 class Images2Neibs(Op):
+
+    __props__ = ("mode",)
+
     def __init__(self, mode='valid'):
         """
         :type mode: str
@@ -33,12 +36,6 @@ class Images2Neibs(Op):
                                       " implemented for the op Images2Neibs")
         self.mode = mode
 
-    def __eq__(self, other):
-        return type(self) == type(other) and self.mode == other.mode
-
-    def __hash__(self):
-        return hash(type(self)) ^ hash(self.mode)
-
     def __str__(self):
         return self.__class__.__name__ + "{%s}" % self.mode
 

theano/tensor/nnet/nnet.py

@@ -44,21 +44,9 @@ class SoftmaxWithBias(gof.Op):
     This L{Op}'s output is softmax(x+b).
     softmax(x[i]) is the i'th distribution over len(x[i]) options.
     """
-
     nin = 2
     nout = 1
-
-    def __init__(self, **kwargs):
-        gof.Op.__init__(self, **kwargs)
-
-    def __eq__(self, other):
-        return type(self) == type(other)
-
-    def __hash__(self):
-        return tensor.hashtype(self)
-
-    def __str__(self):
-        return self.__class__.__name__
+    __props__ = ()
 
     def make_node(self, x, b):
         x = tensor.as_tensor_variable(x)
@@ -284,7 +272,6 @@ class SoftmaxGrad(gof.Op):
     """Gradient wrt x of the Softmax Op"""
     nin = 2
     nout = 1
-
     __props__ = ()
 
     def make_node(self, dy, sm):
@@ -828,19 +815,11 @@ class CrossentropySoftmaxArgmax1HotWithBias(gof.Op):
     """
     nin = 3
     nout = 3
-
+    __props__ = ()
+    
     def __init__(self, **kwargs):
         gof.Op.__init__(self, **kwargs)
 
-    def __eq__(self, other):
-        return type(self) == type(other)
-
-    def __hash__(self):
-        return tensor.hashtype(self)
-
-    def __str__(self):
-        return self.__class__.__name__
-
     def make_node(self, x, b, y_idx):
         x = tensor.as_tensor_variable(x)
         b = tensor.as_tensor_variable(b)
@@ -1058,18 +1037,9 @@ class CrossentropySoftmaxArgmax1HotWithBias(gof.Op):
 class CrossentropySoftmax1HotWithBiasDx(gof.Op):
     nin = 3
     nout = 1
+    __props__ = ()
+    
     """Gradient wrt x of the CrossentropySoftmaxArgmax1HotWithBias Op"""
-    def __init__(self, **kwargs):
-        gof.Op.__init__(self, **kwargs)
-
-    def __eq__(self, other):
-        return type(self) == type(other)
-
-    def __hash__(self):
-        return tensor.hashtype(self)
-
-    def __str__(self):
-        return self.__class__.__name__
 
     def make_node(self, dy, sm, y_idx, **kwargs):
         dy = tensor.as_tensor_variable(dy)
@@ -1269,15 +1239,8 @@ def crossentropy_softmax_max_and_argmax_1hot(x, y_idx, **kwargs):
 
 
 class CrossentropyCategorical1HotGrad(gof.Op):
-
-    def __eq__(self, other):
-        return type(self) == type(other)
-
-    def __hash__(self):
-        return tensor.hashtype(self)
-
-    def __str__(self):
-        return self.__class__.__name__
+    
+    __props__ = ()
 
     def make_node(self, g_y, coding_dist, true_one_of_n):
         return Apply(self, [g_y, coding_dist, true_one_of_n],
@@ -1313,15 +1276,7 @@ class CrossentropyCategorical1Hot(gof.Op):
            away in favour of one with a C implementation.
 
     """
-
-    def __eq__(self, other):
-        return type(self) == type(other)
-
-    def __hash__(self):
-        return tensor.hashtype(self)
-
-    def __str__(self):
-        return self.__class__.__name__
+    __props__ = ()
 
     def make_node(self, coding_dist, true_one_of_n):
         """
@@ -1950,17 +1905,14 @@ from theano import scalar
 
 
 class Prepend_scalar_constant_to_each_row(gof.Op):
+    
+    __props__ = ()
+    
     def __init__(self, val=0):
         if isinstance(val, float):
             val = scalar.constant(val)
         self.val = val
 
-    def __eq__(self, other):
-        return (type(self) == type(other)) and (self.val == other.val)
-
-    def __hash__(self):
-        return tensor.hashtype(self) ^ hash(self.val.data)
-
     def __str__(self):
         return '%s{%s}' % (self.__class__.__name__, self.val)
 
@@ -2007,14 +1959,8 @@ class Prepend_scalar_constant_to_each_row(gof.Op):
 
 
 class Prepend_scalar_to_each_row(gof.Op):
-    def __eq__(self, other):
-        return (type(self) == type(other))
-
-    def __hash__(self):
-        return tensor.hashtype(self)
 
-    def __str__(self):
-        return self.__class__.__name__
+    __props__ = ()
 
     def make_node(self, val, mat):
         # check type of input